The largest Kuiper belt objects

Michael E. BrownCalifornia Institute of Technology

ABSTRACT

While for the first decade of the study of the Kuiper belt, a gap existedbetween the sizes of the relatively small and faint Kuiper belt objects (KBOs)that were being studied and the largest known KBO, Pluto, recent years haveseen that gap filled and the maximum size even expanded. These large KBOsoccupy all dynamical classes of the Kuiper belt with the exception of the coldclassical population, and one large object, Sedna, is the first member of a newmore distant population beyond the Kuiper belt. Like Pluto, most of the largeKBOs are sufficiently bright for detailed physical study, and, like Pluto, mostof the large KBOs have unique dynamical and physical histories that can begleaned from these observations. The four largest known KBOs contain surfacesdominated in methane, but the details of the surface characteristics differ oneach body. One large KBO is the parent body of a giant impact which hasstrewn multiple fragments throughout the Kuiper belt. The large KBOs havea significantly larger satellite fraction than the remainder of the Kuiper belt,including the only known multiple satellite systems and the relatively smallestsatellites known. Based on the completeness of the current surveys, it appearsthat ∼3 more KBOs of the same size range likely still await discovery, but thattens to hundreds more exist in the more distant region where Sedna currentlyresides.

1. Introduction

While once Pluto appeared as a unique objectin the far reaches of the solar system, the discov-ery of the Kuiper belt caused the immediate real-ization that Pluto is a member of a much largerpopulation. But while Pluto’s orbit makes it atypical member of Kuiper belt population dynam-ically, Pluto itself has still remained special as oneof the few trans-Neptunian objects bright enoughfor detailed studies. Much of what we understandof the composition, density, and history of objectsin the Kuiper belt ultimately derives from detailedstudies of Pluto.

Recently, however, surveys of the Kuiper beltbegan to discover Kuiper belt objects (KBOs) ofcomparable and now even larger size than Pluto.The largest survey to date has used the 48-inchPalomar Schmidt telescope to cover almost 20,000

square degrees of sky to a limiting magnitude ofR∼20.5 (Figure 1). This survey has uncoveredmost of the known large KBOs (i.e Trujillo andBrown 2003; Brown et al. 2004, 2005b). A total of71 objects beyond 30 AU have been detected, ofwhich 21 were previously known (or have been in-dependently discovered subsequently). Recoveryof objects is still underway to define the dynamicsof the large objects; to date 54 of the 71 objectshave secure orbits.

This survey for the largest Kuiper belt ob-jects serves both as a search for individual objectsbright enough for detailed study and also as thefirst modern wide-field survey of the outer solarsystem to more fully define the dynamical prop-erties of the entire region. In this chapter we willfirst survey the dynamical properties of the largestKBOs and compare them to the population as awhole, then we will examine the largest individual

1

KBOs, and finally we will review the bulk proper-ties of these largest objects, summarized in Table1.

2. Population properties of the largest

KBOs

2.1. Dynamical distribution

As first noticed by Levison and Stern (2001),KBOs brighter than an absolute magnitude ofabout 6.0 are distributed with a much broader in-clination distribution than those fainter. Physi-cally this trend is better stated that the low in-clination population is missing the largest objects(or at least the brightest objects) that are foundin the high inclination population. This effect iseasily visible in a simple plot of inclination ver-sus absolute magnitude, but such a direct com-parison of simple discovery statistics is thoroughlybiased by the fact that most surveys for fainterKBOs have been restricted much more closely tothe ecliptic and thus preferentially find low incli-nation objects.

One method for examining a population rela-tively unbiased by differences in latitudinal cover-age of surveys is to consider only objects detectedat a restricted range of latitudes. In such debiasedexaminations, no statistically significant differencecan be discerned between the size distribution ofthe high and low inclination populations. Fromthe discovery statistics alone no definitive indica-tion exists that the populations differ. The ques-tion must be addressed with actual measurementsof sky densities of KBOs of different brightnessesrather than simple discovery statistics.

Fortunately, the Palomar survey for large KBOsis complete for low inclinations (with the exceptionof the galactic plane), so we now know that thereare no objects brighter than absolute magnitude4.5 in the low inclination population (or, more per-tinently, in the cold classical region of the Kuiperbelt, defined by Morbidelli and Brown (2005) asthe dynamically and physically distinct subpop-ulation of classical KBOs with uniquely uniformred colors and inclinations lower than about 4 de-grees), while in the excited population (defined asthe resonant, scattered, and hot classical popula-tion) 29 objects brighter than that absolute mag-nitude are currently known to exist, with the cur-rent brightest (Eris) known having an absolute

magnitude of -1.2. The difference in maximumbrightness and presumably maximum size betweenthe cold classical and the excited populations isvast.

This difference in maximum size places a pow-erful constraint on the dynamical rearrangementof the outer solar system. No dynamical processcan preferentially damp the inclinations of onlythe small KBOs nor preferentially excite the incli-nations of only large KBOs, so the high and lowinclination populations must have either formedat different times or in different places. A cur-rent working hypothesis for the larger sizes of thehigh inclination population was suggested by Lev-ison and Stern (2001) and examined in detail byGomes (2003). They noted that the difference insize distribution can be explained if the largest ob-jects formed in the solar nebula closer to the sunwhere nebular densities were higher and growthtimes were faster and that the objects closer tothe sun suffered more extreme scattering by Nep-tune and thus acquired higher inclinations. Otherforces may be at play, however, and a fully con-vincing explanation remains elusive.

A survey of the largest Kuiper belt objects,then, is only a survey of the excited populationsof the Kuiper belt. With this caveat, we cannow examine the spatial distribution of the largestKuiper belt objects. Figure 2 shows the latitudi-nal distribution, corrected for coverage complete-ness, of the KBOs from our survey. The prominentpeaks around 10 degrees north and south eclipticlatitude cannot be modeled with any simple in-clination distribution of objects in circular orbits.Even if all objects in the sky had inclinations of 10degrees or higher, more objects would appear atlower latitudes than are seen in the survey. Whilesuch a latitudinal distribution is impossible for ob-jects with circular or even randomly oriented or-bits, many of the objects are consistent with be-ing resonant objects and thus can have preferen-tial orientations in the sky. Pluto, for example, aswell as many other KBOs in 3:2 resonance withNeptune, comes to perihelion near its maximumexcursion above the ecliptic. This effect will causea magnitude-limited survey to preferentially de-tect resonant objects at large distances above theecliptic. A full examination of this effect awaitsfull dynamical characterization of the survey pop-ulation, but from the preliminary data it appears

2

that resonances are likely able to explain thesehigh latitude concentrations. If true, the high lat-itude concentrations are not likely a characteristicof the largest KBOs, but a general property of thehigh inclination Kuiper belt which has not beenadequately surveyed until now. The resonant pop-ulation may be significantly more populated thanlow latitude surveys have indicated.

2.2. Beyond the Kuiper belt

Among the large objects detected, one appearsdynamically distinct from the entire Kuiper beltpopulation. Sedna has a perihelion well beyondthe main concentration of KBOs and an extremeeccentric orbit with a aphelion around 900 AU(Brown et al. 2004). Though the discovery ofSedna presages a large population in this dis-tant region beyond the Kuiper belt, no surveysfor fainter objects have yet succeeded in detectingsuch distant objects. While some bias against theslow motions of these objects presumably existsin the main KBO surveys, it is also possible thatSedna has an albedo higher than the more numer-ous smaller members of the population. Sednacould be thus, like Pluto, an atypically brightmember of its population which allows us to de-tect it much more easily than would have beenotherwise possible.

Sedna exists in a dynamical region of the solarsystem that was not expected to be occupied. Ithas been proposed to be part of a fossilized innerOort cloud (Brown et al. 2004; Brasser et al. 2006),a product of a single anomalous stellar encounter(Morbidelli and Levison 2004), an object capturedfrom a passing star (Kenyon and Bromley 2004),a consequence of scattering by now ejected Kuiperbelt planets (Gladman and Chan 2006), a signa-ture of perturbation by a distant massive planet(Gomes et al. 2006) and others. Each of theseprocesses creates a dynamically unique popula-tion in this region beyond the Kuiper belt. Find-ing even a handful more of these distant objectsshould give powerful insights into some of the ear-liest processes operating at the beginning of thesolar system.

This distant population could be significantlymore massive than that of the Kuiper belt. Sednais currently near perihelion of its 11,000 yr orbit.It would have been detected in the Palomar surveyonly during a ∼150 year period surrounding peri-

helion, suggesting that the total number of Sedna-sized or larger objects in the distant population isbetween about 40 and 120. The total number ofSedna-sized or larger objects in the Kuiper belt is∼ 5 − 8. If the distant population has the samesize distribution as the Kuiper belt – which seemslikely given that the Kuiper belt is the most likelysource region for this population – this numberof Sedna-sized objects suggests a total mass atleast an order of magnitude higher than that inthe Kuiper belt.

2.3. Size distribution

A finite discrete population which generally fol-lows a power-law size distribution cannot maintainthis distribution at the largest sizes. Early surveysof the Kuiper belt expected that for the brightestobjects the number of detections would fall sig-nificantly below the power-law prediction. Figure3 shows the opposite. For objects brighter thanR ∼ 19.8 the power law found by Bernstein et al.(2004) for the excited population falls well short ofthe actual numbers of detections. This increase inthe numbers of bright objects over that expectedis a consequence of the general increase in albedowith size occurring for these objects (see chapterby Stansberry et al.). A plot of number of ob-jects versus absolute magnitude shows the sametrend (with a bias towards higher absolute mag-nitude because of the flux-limited nature of thesurvey), and the location of the deviation fromthe power law is a useful indicator of the approx-imate location where albedo changes begin to beimportant. Eight objects brighter than H ∼ 3 de-viate most strongly from the power law and are aconvenient dividing line between the largest indi-vidual KBOs and the remaining population. Eachof these largest KBOs has interesting unique prop-erties that we describe below.

3. Individual properties of the largest

KBOs

3.1. Eris

Eris is currently the largest known object in theKuiper belt. Direct measurement of the size withthe Hubble Space Telescope gives a diameter of2400± 100 km (Brown et al. 2006a), while radio-metric measurement with IRAM gives 3000± 400(Bertoldi et al. 2006). While the two measure-

3

ments appear discrepant, they only differ by 1.5σ

owing to the large uncertainty in the radiomet-ric measurement. We will take the measurementwith the smaller uncertainty for the remainder ofthe discussion but comment on the larger diame-ter at the end. This size measurement implies aremarkably high V-band albedo of 0.86± 0.07.

The infrared spectrum of Eris is dominated byabsorption from methane similar to the spectrumof Pluto (Brown et al. 2005b) (Figure 4). Un-like Pluto, however, the infrared spectrum of Erisshows no evidence for the small shifts in the wave-length of the methane absorption associated withmethane being dissolved in a nitrogen matrix. Theweakest methane absorptions in the visible, how-ever, do possibly show a small shift (Licandro et al.2006a), perhaps suggesting that methane and ni-trogen are layered, with mostly pure nitrogen onthe surface (where it is probed by the strong ab-sorption features which show no shifts) and dis-solved nitrogen below (where it is probed by theweak absorption features which require long pathlengths to appear). The weak 2.15 µm absorptionfeature of nitrogen ice has not been definitivelyidentified, but at the low temperature expectedon Eris nitrogen should be in its α, rather than β

form as it is on Pluto. The α form has an absorp-tion even weaker than that of the β form (Grundyet al. 1993; Tryka et al. 1995) so detection may beextremely difficult even if nitrogen is indeed abun-dant.

The visible properties of Eris also differ fromthose of Pluto. Eris is less red than Pluto, and,while Pluto has one of the highest contrast sur-faces in the solar system and varies in brightnessby 36% over a single rotation (see Brown 2002), novariation has been seen on Eris to an upper limitof 0.05 magnitudes (Carraro et al. (2006) reporta photometric variation on one of five nights of∼0.02 magnitudes, but no additional observationshave confirmed this potential long-term variabil-ity).

The high albedo, lower red coloring, and lackof rotational variation on Eris are all consistentwith a surface dominated by seasonal atmosphericcycling. With Eris currently near aphelion at 97AU the radiative equilibrium temperature is ∼ 20K and nitrogen and methane have essentially zerovapor pressures, compared to vapor pressures of 17µbar and 2 nbar at the ∼36 K equilibrium tem-

perature at the 38 AU perihelion. At the cur-rent aphelion position, the perihelion atmosphereshould be collapsed onto the surface as 0.6 µmof methane and 2 mm of nitrogen. The darkerand redder regions such as those on Pluto, whichgive Pluto its strong contrast, red color, and loweraverage albedo, should be covered, giving Eris amore uniform, brighter, and less red surface. In-deed, the high albedo of Eris appears similar toindividual regions on Pluto where no dark mate-rial appears to be present (Young et al. 2001). AsEris proceeds from aphelion and the surface warmswe should expect that darker regions will becomeuncovered and the surface will appear darker, red-der, and more Pluto-like. While this story forthe seasonal evolution of surface of Eris consis-tently explains many aspects of the observations,the pure methane ice on the surface remains un-explained. Methane will freeze out before nitro-gen, so the surface might be expected to be lay-ered with methane below nitrogen, with perhaps amixed Pluto-like layer from perihelion below, butbetter constrained observations and more detailedmodeling will be required to understand the sur-face state and evolution.

Eris is orbited by an apparently single satel-lite, Dysnomia, approximately 250 times fainterthan Eris (Brown et al. 2006b; Brown and Schaller2007). More distant satellites up to 10 timesfainter than Dysnomia can be ruled out fromdeep HST observations (Brown and Suer 2007).Models for satellite capture which appear success-ful at describing many of the large satellites de-tected around many Kuiper belt objects (Goldre-ich et al. 2002) cannot account for the presenceof such a small satellite. The most likely cre-ation mechanism appears to be impacts such asmodeled by Canup (2005) who, while attemptingto find models describing the Charon-forming im-pact, found many cases in which the impact gen-erated a disk which could coalesce to form a muchsmaller satellite. The near circular orbit of Dysno-mia (e < 0.017) is also consistent with the idea offormation from a disk and outward tidal evolution.

From the orbit of Dysnomia, the mass of Erisis found to be (1.66 ± 0.02) × 1022 kg or 27 ± 2%greater than that of Pluto (Brown and Schaller2007). Using the HST size measurement the den-sity is then 2.3±0.3 g cm−3, with almost all of theuncertainty due to the uncertainty in the size mea-

4

surement. The density is consistent on the lowerend with the 2.03 ± 0.06 g cm−3 density of Plutoand on the high end with the ∼2.6 g cm−3 densityof 2003 EL61 (see below). Note that the largersize measurement from IRAM would give a den-sity of 1.1±0.6 g cm−3, which, when comparing toother large KBOs and icy satellites satellites, ap-pears unreasonably low for an object of this size(see below).

We might expect that the disk forming impactthat generated Dysnomia would have removedsome ice from Eris, leading to a higher densitythan that of Pluto. A more accurate measurementof the density, which would require a more accu-rate measurement of the size, is clearly warranted.It appears that only an occultation is likely to givean improved size estimate for Eris, and, with Erisfar from the galactic plane, opportunities will belimited.

3.2. Pluto

Pluto, discussed in detail in the chapter byStern, is the largest object in the highly populated3:2 mean motion resonance with Neptune. Its highalbedo and current position near perihelion makeit the brightest object in the Kuiper belt and thusthe first discovered and most heavily observed.Physically, it appears to be a slightly smaller twinof Eris. The main visible differences appear to bethe redder color, the presence of dark areas on thesurface, and the different state of methane on thesurface. As discussed above, most of these differ-ences can be explained as an expected consequenceof the closer heliocentric distance of Pluto.

Pluto is surrounded by a system of one large(Charon)(Christy and Harrington 1978) and twosmall satellites (Nix and Hydra)(Weaver et al.2006). Modeling by Canup (2005) suggests thatthe large satellite Charon can be explained asa consequence of a grazing collision between theproto-Pluto and Charon in which little exchangeor heating takes place. While no detailed modelingof the formation of the smaller satellites has beenperformed, their similar orbital plane to Charonand near-circular orbits (Buie et al. 2006) suggestthat they were formed in the same collision.

3.3. Sedna

While the size of Sedna remains uncertain, anupper limit can be placed from Spitzer observa-tions (see chapter by Stansberry et al.), and amore tenuous lower limit can be placed by assum-ing that the geometric albedo at all wavelengthsis lower than 100% (which need not necessarilybe true). These limits constrain the V albedo ofSedna to be between 0.16 and 0.30 and the diame-ter to be between 1200 and 1600 km. A deep HSTsearch for satellites has revealed no candidates toa limit of about 500 times fainter than the primary(Brown and Suer 2007).

Sedna is one of the reddest KBOs known, and inmoderate signal-to-noise data, the infrared spec-trum appears to contain methane and perhapsnitrogen (Barucci et al. 2005)(Figure 4). Thevisible-to- infrared spectrum and moderate albedois consistent with an object covered in dark red or-ganic tholins but with some covering of methaneand nitrogen frosts. Sedna is currently at 90 AUand 70 years away from its 76 AU perihelion inits 11,000 year orbit which takes it to 900 AU.It is currently warming and developing whateverlimited atmosphere it will have. A 76 AU equi-librium temperature atmosphere of ∼160 nbar ofnitrogen will correspond to a ∼40 µm solid layerof nitrogen ice at aphelion and a ∼36 µm layer atits current position of 90 AU.

The darker and redder surface of Sedna is con-sistent in albedo and color with the darker regionson Pluto. The long orbital period and high ec-centricity mean that Sedna spends very little timenear perihelion, so much more time is available forsolid state processing of the material than there isfor surface regeneration. The extremely low tem-perature of Sedna prevents much of an atmosphereeven near perihelion and thus no extensive frostsurface should ever develop.

3.4. 2005 FY9

2005 FY9 is the brightest KBO after Pluto,and radiometric measurements from the SpitzerSpace Telescope (see chapter by Stansberry etal.) suggest a diameter of 1500 ± 300 and analbedo of 80+10

−20%. Like Eris, Pluto, and Sedna,2005 FY9 has a surface spectrum dominated bymethane (Barkume et al. 2005; Licandro et al.2006b; Brown et al. 2007a), but the methane ab-

5

sorption features on 2005 FY9 are significantlydeeper and broader than those on the other ob-jects (Fig. 4). The depth and breadth of solidstate absorption features is a function of opticalpath length through the absorbing material, so thefeatures on 2005 FY9 can be interpreted as beingdue to extremely large (∼ 1 cm) methane grainson 2005 FY9, or, likely more appropriately, as dueto a slab of methane ice with scattering impuritiesseparated by ∼1 cm. Methane grain sizes on theother bodies are closer to 100µm in contrast.

In addition to the large methane path lengths,2005 FY9 differs from Pluto in that even mod-erately high signal-to-noise spectra show no ev-idence for the presence of the 2.15 µm nitrogenice absorption feature (Brown et al. 2007a). Ni-trogen appears depleted on 2005 FY9 relative tomethane by at least an order of magnitude com-pared to Pluto. Visible spectroscopy shows evi-dence, however, for slight shifts in the wavelengthsof the methane absorptions features which couldbe indicative of a small amount of surface coverageof methane dissolved inside nitrogen (Tegler et al.2007).

Finally, 2005 FY9 has a clear signature of thepresence of small grains of ethane, in addition tothe methane (Brown et al. 2007a). Ethane is oneof expected dissociation products of both gaseousand solid-state methane.

All of these unique characteristics of 2005 FY9can be interpreted as being due to a large de-pletion of nitrogen on the object. The depletionof nitrogen would make methane the dominantvolatile on the surface and allow grains of rela-tively pure methane to grow large as the grains ofnitrogen do on Pluto. In addition, the presence ofmethane in pure rather than diluted form wouldallow the solid state degradation of methane toethane that would not be possible with methanediluted in small concentrations in nitrogen. 2005FY9 may be a transition between the larger sur-face volatile-rich objects and the smaller surfacevolatile-depleted objects.

2005 FY9 is the largest KBO to have no knownsatellite. Deep observations from HST place anupper limit for the brightness of faint distant satel-lites of one part in 10000 (Brown and Suer 2007).

3.5. 2003 EL61

2003 EL61 was first found to be unusual due toits rapid rotation and large light curve variation.Rabinowitz et al. (2006) inferred that 2003 EL61was a rapidly rotating ellipsoid with a 4 hour ro-tation period. Assuming that the primary spinsin the same plane as the first satellite discovered(Brown et al. 2005a), the light curve and periodsuggest a body with a density of 2.6 g cm−3, asize (based on the density and mass determinedfrom the satellite orbit) of 2000 x 1500 x 1000 km,and a visual albedo (based on the derived size andon the brightness) of 0.73 (the formal uncertain-ties on these parameters are small, but probablydo not reflect the true uncertainties in our under-standing of the interior state of large icy bodiesand the degree to which uniform denisty hydro-static equilibrium holds), consistent with infraredspectra showing deep water ice absorption (Tru-jillo et al. 2007).

Infrared spectroscopy of the satellite revealedthe deepest water ice absorption features of anybody detected in the outer solar system(Barkumeet al. 2006), which effectively ruled out a captureorigin, as capture of a spectrally unique body ap-pears implausible. The rapid rotation, high den-sity, unusual satellite spectrum, and the discoveryof a second inner satellite (Brown et al. 2006b) allstrongly point to a collisional origin for this sys-tem.

A large infrared survey showed that a smallnumber of KBOs have deep water ice absorptionssimilar to that of 2003 EL61 and almost as deepas its satellite (see Figure 4 in chapter by Barucciet al.). Remarkably, these KBOs are all dynami-cally clustered near the dynamical position of 2003EL61 itself. Determination of the proper orbitalelements of these objects shows that they repre-sent a tight dynamical family separated by only140 m s−1 (Brown et al. 2007b). Such a tightdynamical clustering is itself unusual enough; cou-pled with the spectral similarity and the additionalevidence for a giant impact, it becomes clear thatthe objects in this family are the collisional frag-ments of a giant impact on the proto-2003 EL61.While the fragments themselves are tightly clus-tered, 2003 EL61 itself has a velocity differenceof approximately 500 m s−1 from the fragments.This difference is easily explained by the residence

6

of 2003 EL61 with the 12:7 mean motion resonancewith Neptune which causes long term eccentricityand inclination evolution that can take an objectfrom near the center of the cluster to the positionof 2003 EL61 on a time scale of ∼1 Gyr.

While a giant impact on the proto-2003 EL61appears capable of explaining each of the individ-ual observations, some mysteries remain. In mod-eling to date, impacts are seen to either dispersefragments or create a disk out of which satellitescan form. 2003 EL61 appears to have done both.In addition, the very small velocity dispersion ofthe family implies that the fragments left the sur-face of 2003 EL61 with velocities a small fractionabove the 1 km s−1 escape velocity. Detailed mod-eling will be required for a further understandingof the 2003 EL61 system.

3.6. Other large objects

The three other objects in our collection of largeKBOs each also have unique properties. Quaoarand Orcus each have water ice absorption amongthe deepest of non-2003 EL61 fragment KBOs (Je-witt and Luu 2004; de Bergh et al. 2005; Tru-jillo et al. 2005). Ixion is the largest known ob-ject with a nearly featureless infrared spectrum(Brown et al. 2007b).

The infrared spectrum of Quaoar has an ab-sorption feature at 2.2 µm that has been in-terpreted as being due to ammonia (Jewitt andLuu 2004) in analogy to an absorption feature onCharon (Brown and Calvin 2000), though the twospectra appear different. The absorption feature isalso, however, consistent with the position of oneof the strongest absorptions for methane. Moredetailed observations to constrain the compositionof the surface of Quaoar are clearly warranted.Quaoar is, in addition, the smallest object knownto have a faint satellite (fractional brightness of0.6%) like those of Eris, Pluto, and 2003 EL61(Brown and Suer 2007).

Orcus is a Plutino with an orbit which is nearlya mirror-image of that of Pluto. It is the largestKBO with an (apparently) single large (fractionalbrightness of 8%) satellite; deep HST images showthat any more distant satellites must be fainterthan Orcus by at least a factor of 1000 (Brown andSuer 2007). Outer satellites of the relative faint-ness of those of Pluto would remain undetected.

The satellite of Orcus is on a near-circular orbitwith a 9.5 day period, consistent with outwardevolution from an initially tighter orbit.

Ixion is the brightest object in absolute magni-tude with a nearly featureless infrared spectrum,though it is not clear that it is the largest such ob-ject. Spitzer observations (see chapter by Stans-berry et al.) only moderately constrain the sizeto 590±190 km and the albedo to between ∼9and 30%. A handful of other KBOs have Spitzermeasurements of a similar or greater size, includ-ing Varuna, Huya, 2002 AW197, 2002 UX25, 2004GV9, 2002 MS4 and 2003 AZ84, and their derivedalbedos range from 6 to 30%. Some of these ob-jects (2002 UX25 and 2003 AZ84) are known tohave moderately large close satellites, and one –Varuna – is known to be a rapid rotator with sim-ilarities to 2003 EL61 (Jewitt and Sheppard 2002),but, in general, these objects appear to share fewof the properties of the unique larger KBOs.

4. Ensemble properties

4.1. Surface composition

The most striking visible difference between thelargest KBOs and the remainder of the popula-tion is the presence of volatiles such as methane,nitrogen, and CO in the spectra of the the largeobjects compared to relatively featureless spectraof the remaining objects. The transition fromsmall objects with volatile-free to large objectswith volatile-rich surfaces appears to be explain-able with a simple model of atmospheric escapeshown in Figure 5 (Schaller and Brown 2007).Most KBOs are too small and too hot to be ableto retain volatiles against atmospheric escape overthe life of the solar system, a few objects are solarge or so cold that they easily retain volatiles,and a small number are in the potential transitionregion between volatile free and volatile rich sur-faces. 2003 EL61 is sufficiently large that it couldretain volatiles, but it seems likely that the giantimpact which removed much of its water ice wouldhave removed much of the volatile mass, also, ei-ther through direct ejection or heating. 2005 FY9and Quaoar are both sufficiently hot that the low-vapor-pressure nitrogen should all have escaped,but the lower-vapor-pressure methane could stillbe retained. This depletion of nitrogen relativeto methane is precisely what is observed on 2005

7

FY9. On Quaoar, if the 2.2 µm absorption is in-terpreted as being due to methane instead of am-monia, it would appear that Quaoar is in the laststages of volatile loss.

The model shown in Figure 5 provides thefirst basic framework for understanding the sur-face compositions of the objects in the Kuiperbelt. The vast majority of the known objects aretoo small and/or too hot to have the possibilityof retaining any surface volatiles. Surfaces dom-inated by relatively featureless involatile heavierorganics or exposures of water ice (see chapter byBarucci et al.) are therefore expected on such ob-jects. Volatile-rich surfaces are only possible onthese largest of the bodies in the Kuiper belt. Inthe region beyond the Kuiper belt inhabited bybodies such as Sedna, we should expect that mostof the bodies – even relatively small ones – willhave the capability of retaining surface volatiles.

The largest non-methane objects have the deep-est water ice absorption features (ignoring the pre-sumably special case of 2003 EL61 and its frag-ments), even taking into account the lower signal-to-noise of the spectra of the fainter objects (seeFigure 4 of the chapter by Barucci et al.). Unlikefor the presence or absence of surface volatiles noclear explanation of this trend is apparent, thougha partial explanation could include the initiallyhigher temperatures of the larger objects leadingto greater internal volatile loss and perhaps dif-ferentiation. Fewer organic volatiles could thenlead to less creation of dark organic tholins. Sucha process would lead to higher albedos for theselarger objects, which is indeed observed, but alsobluer colors, which is not observed. An alternativeexplanation could invoke the satellite-forming im-pacts that these objects experience in an attemptto explain their surface compositions. Our under-standing of the processes affecting the colors andcompositions of all of the objects in the Kuiperbelt is still primitive.

4.2. Satellites

The largest KBOs appear to have a differentstyle of satellite formation than the other objects.These objects have a greater frequency of satel-lites, the only two known multiple satellite sys-tems, and the possibility of much smaller satellites.Brown et al. (2006b) found in an adaptive opticssurvey of the four largest KBOs that the proba-

bility that three out of four of these would havedetectable satellites suggests that they are drawnfrom a different population than the remainder ofthe Kuiper belt at the 98.2% confidence level. Up-dating this calculation for our currently definedpopulation, we find that the probability that fiveor more out of eight in our sample of large ob-jects are drawn from the same population as theremainder of the Kuiper belt is less than 1%.

The presence of relatively small satellitesaround Eris, Pluto, 2003 EL61, and Quaoar sug-gests formation by impact, rather than dynamical-friction aided capture. The moderate size andtight circular orbit of the satellite of Orcus couldalso indicate a collisional rather than capture ori-gin. After the early discovery of near-equal bright-ness well-separated eccentrically-orbiting KBO bi-naries (see chapter by Noll et al.) much empha-sis was placed on trying to explain the genesisof these unusual systems through some sort ofcapture mechanism. Collisions, however, appeara dominant satellite-creating process among thelargest KBOs and perhaps also for the now nu-merous known closely spaced binaries.

4.3. Densities

The abundance of satellites and the ability tomake accurate size measurements (see chapter byStansberry et al.) allows determination of the den-sity for many of the largest KBOs. While thehandful of smaller KBOs with known densities ap-pear to have unexpectedly low densities of ∼ 1g cm−3 and even lower (Stansberry et al. 2006),the largest KBOs have densities between ∼ 1.9and 2.5 g cm−3 as expected from cosmochemicalabundances in the outer solar system (McKinnonand Mueller 1989). Figure 6 shows the measureddensities of the large KBOs, including Triton andCharon. Within the largest KBOs, no statisti-cally significant trend exists in KBO density ver-sus radius. Similarly, no significant trend is seenin the densities of the icy satellites of the outerthree planets through this size range. A rank cor-relation test shows that the KBOs are more densethan the icy satellites, however, at the 95% con-fidence level. These higher densities could be theresult of the different formation environment be-tween the proto-solar and proto-giant planet neb-ulae, though a bias in KBO densities caused byimpacts cannot be ruled out, as density measure-

8

ments of KBOs (with the exception of Triton) re-quire the presence of a satellite.

5. Conclusions

Each of the largest KBOs has a unique dynami-cal and physical history which can be gleaned fromdetailed observations such as described here. As awhole, the largest KBOs appear distinct in surfacecomposition, satellite frequency and style, anddensity. Impacts appear to have played a more dis-cernible role among the largest KBOs than amongthe population at large.

Based on the latitudinal completeness of thePalomar survey, it appears that 2 or 3 moreKBOs of the size range of those described herelikely await discovery, though many more large ob-jects must exist in the distant regions beyond theKuiper belt. The most likely location to find largeundiscovered Kuiper belt objects is in the bandat 10 degrees south ecliptic latitude where the skydensities are highest and the completeness is low-est, though with the low numbers remaining to befound they could be almost anywhere within theKuiper belt.

Several outstanding questions remain about thelargest Kuiper belt objects:

• Why are there no large KBOs among thecold classical population?

• What does Sedna’s dynamical location tellus about the history of the solar system?

• What causes the density enhancements at±10 degrees ecliptic latitude and what im-plication does this have for the formation ofthe Kuiper belt?

• How does atmospheric cycling affect thepresence and layering of species on volatile-rich large KBOs?

• Why are the water absorption features of2003 EL61 and its satellites and fragmentsdistinctly deeper than those of other water-rich KBOs?

• Are multiple satellite systems commonamong the large KBOs?

• Is Quaoar at the transition from having avolatile-rich to a volatile-poor surface?

• Are any active sources of methane, such asserpentinization of ultramafic rock, neces-sary to explain the volatiles on the largestKBOs?

• Is the impact frequency required to explainall of the presumably impact-related featuresof the large KBOs higher than expected?

• Do impacts such as those experienced by2003 EL61 raise the densities on otherKBOs?

• What explains the difference between thewater ice-rich surfaces of some moderatesized KBOs and the spectrally featurelesssurfaces of others?

The recent discoveries of these largest KBO en-sures an accessible population for addressing thesequestions and promises a slew of new questions asmore details of these objects are discerned.

REFERENCES

Barkume, K. M., M. E. Brown, and E. L. Schaller.2005. Near Infrared Spectroscopy of Icy Plan-etoids. AAS/Division for Planetary SciencesMeeting Abstracts 37:52.11.

Barkume, K. M., M. E. Brown, and E. L. Schaller.2006. Water Ice on the Satellite of Kuiper BeltObject 2003 EL61. ApJ640:L87–L89.

Barucci, M. A., D. P. Cruikshank, E. Dotto,F. Merlin, F. Poulet, C. Dalle Ore, S. Fornasier,and C. de Bergh. 2005. Is Sedna another Tri-ton? A&A439:L1–L4.

Bernstein, G. M., D. E. Trilling, R. L. Allen, M. E.Brown, M. Holman, and R. Malhotra. 2004.The Size Distribution of Trans-Neptunian Bod-ies. AJ128:1364–1390.

Bertoldi, F., W. Altenhoff, A. Weiss, K. M.Menten, and C. Thum. 2006. The trans-neptunian object UB313 is larger than Pluto.Nature439:563–564.

Brasser, R., M. J. Duncan, and H. F. Levison.2006. Embedded star clusters and the forma-tion of the Oort Cloud. Icarus 184:59–82.

9

Brown, M. E. 2002. Pluto and Charon: Forma-tion, Seasons, Composition. Annual Review ofEarth and Planetary Sciences 30:307–345.

Brown, M. E., K. M. Barkume, G. A. Blake, E. L.Schaller, D. L. Rabinowitz, H. G. Roe, andC. A. Trujillo. 2007a. Methane and Ethaneon the Bright Kuiper Belt Object 2005 FY9.AJ133:284–289.

Brown, M. E., K. M. Barkume, D. Ragozzine, andE. L. Schaller. 2007b. Discovery of an Icy Col-lisional Family in the Kuiper Belt. Nature, inpress .

Brown, M. E., A. H. Bouchez, D. Rabinowitz,R. Sari, C. A. Trujillo, M. van Dam, R. Camp-bell, J. Chin, S. Hartman, E. Johansson, R. La-fon, D. Le Mignant, P. Stomski, D. Summers,and P. Wizinowich. 2005a. Keck ObservatoryLaser Guide Star Adaptive Optics Discoveryand Characterization of a Satellite to the LargeKuiper Belt Object 2003 EL61. ApJ632:L45–L48.

Brown, M. E., and W. M. Calvin. 2000. Evi-dence for Crystalline Water and Ammonia Iceson Pluto’s Satellite Charon. Science 287:107–109.

Brown, M. E., and E. L. Schaller. 2007. The massof dwarf planet Eris. submitted.

Brown, M. E., E. L. Schaller, H. G. Roe, D. L.Rabinowitz, and C. A. Trujillo. 2006a. DirectMeasurement of the Size of 2003 UB313 fromthe Hubble Space Telescope. ApJ643:L61–L63.

Brown, M. E., and T.-A. Suer. 2007. Small satel-lites of big Kuiper belt objects. AJ, submitted.

Brown, M. E., C. Trujillo, and D. Rabinowitz.2004. Discovery of a Candidate Inner OortCloud Planetoid. ApJ617:645–649.

Brown, M. E., C. A. Trujillo, and D. L. Ra-binowitz. 2005b. Discovery of a Planetary-sized Object in the Scattered Kuiper Belt.ApJ635:L97–L100.

Brown, M. E., M. A. van Dam, A. H. Bouchez,D. Le Mignant, R. D. Campbell, J. C. Y. Chin,A. Conrad, S. K. Hartman, E. M. Johansson,R. E. Lafon, D. L. Rabinowitz, P. J. Stomski,

Jr., D. M. Summers, C. A. Trujillo, and P. L.Wizinowich. 2006b. Satellites of the LargestKuiper Belt Objects. ApJ639:L43–L46.

Buie, M. W., W. M. Grundy, E. F. Young, L. A.Young, and S. A. Stern. 2006. Orbits and Pho-tometry of Pluto’s Satellites: Charon, S/2005P1, and S/2005 P2. AJ132:290–298.

Canup, R. M. 2005. A Giant Impact Origin ofPluto-Charon. Science 307:546–550.

Carraro, G., M. Maris, D. Bertin, and M. G.Parisi. 2006. Time series photometryof the dwarf planet ERIS (2003 UB313).A&A460:L39–L42.

Christy, J. W., and R. S. Harrington. 1978. Thesatellite of Pluto. AJ83:1005–+.

de Bergh, C., A. Delsanti, G. P. Tozzi, E. Dotto,A. Doressoundiram, and M. A. Barucci. 2005.The surface of the transneptunian object 90482Orcus. A&A437:1115–1120.

Gladman, B., and C. Chan. 2006. Production ofthe Extended Scattered Disk by Rogue Planets.ApJ643:L135–L138.

Goldreich, P., Y. Lithwick, and R. Sari. 2002.Formation of Kuiper-belt binaries by dy-namical friction and three-body encounters.Nature420:643–646.

Gomes, R. 2003. The Common Origin of the HighInclination TNO’s. Earth Moon and Planets92:29–42.

Gomes, R. S., J. J. Matese, and J. J. Lissauer.2006. A distant planetary-mass solar compan-ion may have produced distant detached ob-jects. Icarus 184:589–601.

Grundy, W. M., B. Schmitt, and E. Quirico.1993. The Temperature-Dependent Spectra ofα and β Nitrogen Ice with Application to Tri-ton. Icarus 105:254–258.

Jewitt, D. C., and J. Luu. 2004. Crystalline waterice on the Kuiper belt object (50000) Quaoar.Nature432:731–733.

Jewitt, D. C., and S. S. Sheppard. 2002. PhysicalProperties of Trans-Neptunian Object (20000)Varuna. AJ123:2110–2120.

10

Kenyon, S. J., and B. C. Bromley. 2004. Stel-lar encounters as the origin of distant So-lar System objects in highly eccentric orbits.Nature432:598–602.

Levison, H. F., and S. A. Stern. 2001. On the SizeDependence of the Inclination Distribution ofthe Main Kuiper Belt. AJ121:1730–1735.

Licandro, J., W. M. Grundy, N. Pinilla-Alonso,and P. Leisy. 2006a. Visible spectroscopy of2003 UB313: evidence for N2 ice on the surfaceof the largest TNO? A&A458:L5–L8.

Licandro, J., N. Pinilla-Alonso, M. Pedani,E. Oliva, G. P. Tozzi, and W. M. Grundy.2006b. The methane ice rich surface of largeTNO 2005 FY9: a Pluto-twin in the trans-neptunian belt? A&A445:L35–L38.

McKinnon, W. B., and S. Mueller. 1989.The density of Triton - A prediction. Geo-phys. Res. Lett.16:591–594.

Morbidelli, A., and M. E. Brown. 2005. CometsII, ed. M.C. Festou, H.U. Keller, H.A. Weaver(U. Arizona Press: 2005).

Morbidelli, A., and H. F. Levison. 2004. Sce-narios for the Origin of the Orbits of the Trans-Neptunian Objects 2000 CR105 and 2003 VB12(Sedna). AJ128:2564–2576.

Rabinowitz, D. L., K. Barkume, M. E. Brown,H. Roe, M. Schwartz, S. Tourtellotte, andC. Trujillo. 2006. Photometric ObservationsConstraining the Size, Shape, and Albedo of2003 EL61, a Rapidly Rotating, Pluto-sizedObject in the Kuiper Belt. ApJ639:1238–1251.

Schaller, E. L., and M. E. Brown. 2007. VolatileLoss and Retention on Kuiper Belt Objects.ApJ, in press .

Stansberry, J. A., W. M. Grundy, J. L. Margot,D. P. Cruikshank, J. P. Emery, G. H. Rieke,and D. E. Trilling. 2006. The Albedo, Size, andDensity of Binary Kuiper Belt Object (47171)1999 TC36. ApJ643:556–566.

Tegler, S. C., W. M. Grundy, W. Romanishin,G. J. Consolmagno, K. Mogren, and F. Vilas.2007. Optical Spectroscopy of the Large KuiperBelt Objects 136472 (2005 FY9) and 136108(2003 EL61). AJ133:526–530.

Trujillo, C. A., and M. E. Brown. 2003. The Cal-tech Wide Area Sky Survey. Earth Moon andPlanets 92:99–112.

Trujillo, C. A., M. E. Brown, K. M. Barkume,E. L. Schaller, and D. L. Rabinowitz. 2007.The Surface of 2003 EL61 in the Near-Infrared.ApJ655:1172–1178.

Trujillo, C. A., M. E. Brown, D. L. Rabinowitz,and T. R. Geballe. 2005. Near-Infrared Sur-face Properties of the Two Intrinsically Bright-est Minor Planets: (90377) Sedna and (90482)Orcus. ApJ627:1057–1065.

Tryka, K. A., R. H. Brown, and V. Anicich. 1995.Near-infrared absorption coefficients of solid ni-trogen as a function of temperature. Icarus116:409–414.

Weaver, H. A., S. A. Stern, M. J. Mutchler,A. J. Steffl, M. W. Buie, W. J. Merline,J. R. Spencer, E. F. Young, and L. A. Young.2006. Discovery of two new satellites of Pluto.Nature439:943–945.

Young, E. F., R. P. Binzel, and K. Crane. 2001. ATwo-Color Map of Pluto’s Sub-Charon Hemi-sphere. AJ121:552–561.

This 2-column preprint was prepared with the AAS LATEX

macros v5.0.

11

Fig. 1.— Coverage of the Palomar survey for large Kuiper belt objects. The map is centered at RA anddeclination of 0 degrees. The white line shows the ecliptic. Approximately 20,000 square degrees north of-30 degrees declination, mostly avoiding the galactic plane, have been covered to a limiting magnitude ofR∼20.5. 71 large KBOs have been found in the survey, including most of the large KBOs discussed here.

12

Fig. 2.— The latitudinal distribution of objectsfound in the Palomar survey for large KBOs. Thelower line with dots shows the number of KBO de-tections in two degree bins. The dashed line showsthe fractional sky coverage as a function of eclip-tic latitude. Sky coverage is incomplete becauseof galactic plane avoidance, (substantial) gaps be-tween CCDs in the mosaic camera, and occasionallack of sky coverage. The solid line above the dotsshows the expected number of large KBOs per lat-itude bin corrected for sky coverage. The promi-nent peaks in sky density at -10 and +10 eclipticlatitude are likely a general property of the highinclination Kuiper belt rather than a property ofonly the large KBOs.

Fig. 3.— The cumulative magnitude distribu-tion of the large KBOs found in the Palomar sur-vey. The upper plot shows the total number ofKBOs detected brighter than a limiting R magni-tude, while the straight line shows the slope of theBerstein et al. (2004) power law fit to the distribu-tion of the excited population. The deviation fromthe power law at magnitudes fainter than ∼20.5 isan indication of where the survey begins to be-come incomplete. The deviation from the powerlaw for the brightest objects, also seen in the dis-tribution of absolute magnitude in the lower plot,is an indication of the increase in albedo of thelargest objects.

13

Fig. 4.— Visible-to-infrared spectra of the fourmethane-covered objects (Barucci et al. 2005;Brown et al. 2007a; Brown 2002; Brown et al.2005b). While each of the objects is dominatedby the signature of methane (with the exceptionof Sedna where the signal is weak but convinc-ing), major differences appear in the objects’ sur-face compositions. Methane on Eris and 2005 FY9appears to be dominantly in pure form, while onPluto much of the methane is dissolved in N2,whose spectral signature can be seen at 2.15 µm.On 2005 FY9, large path lengths through puremethane give rise to broad saturated bands, andabsorption due to ethane can be seen at whatshould be the flat bottom of the 2.3 µm methaneabsorption. The low signal-to-noise of the Sednaspectra prevents detailed analysis , but the weak-ness of the methane and the possible presence of abroad N2 line show a different surface character.

500 1000 1500 2000 2500Diameter (km)

20

30

40

50

Equi

vale

nt T

empe

ratu

re (K

)Sedna

Eris

CharonOrcus

2003 EL612005 FY9Quaoar

PlutoTriton

All Volatiles Lost Volatiles Retained

CH4CO N2

2004 XR190

2000 CR105

Fig. 5.— A model of surface volatile loss onobjects in the Kuiper belt (Schaller and Brown,2007). Most objects in the Kuiper belt are suffi-ciently small or sufficiently hot that atmosphericloss will remove all accessible surface volatiles overthe lifetime of the solar system. No volatiles havebeen detected on any of these objects. A smallnumber of objects are large enough or cold enoughto easily retain surface volatiles, and each of thesehas indeed had surface volatiles detected. Threeobjects are in the transition region between cer-tain volatile loss and possible volatile retention.2003 EL61 has no volatiles detected on the surface,but the mantle shattering impact that it likelyexperienced would likely have removed many ofthe volatiles along with much of the water ice.2005 FY9 indeed appears to be a transition ob-ject as the model predicts, with methane clearlypresent, but a large depletion of nitrogen relativeto methane. Quaoar has a dominantly water icespectra, but an absorption feature at 2.2µm couldbe interpreted as being due to the strongest bandof methane being weakly present, implying thatQuaoar, too, is currently undergoing the transi-tion from a volatile-rich to volatile-poor surface.

14

Fig. 6.— Densities of the largest KBOs, shownwith solid circles. No clear trend exists in densitywith size, though several KBOs with smaller sizesare known to have significantly lower densities. Nostatistically significant trend is seen among thedensities of icy satellite of the outer three plan-ets over this same size range (open triangles; thetriangle with an arrow represents Titan, with a di-ameter of 5150 km). The KBOs are more densethan the satellites at the 95% confidence level.

15

Table 1

Properties of the largest Kuiper belt objects

Eris Pluto 2005 FY9 2003 EL61 Sedna Quaoar Orcus Ixion

diameter (km) 2400±100 2290 1500±300 ∼2000 x 1500 x 1000 1300-1800 1260±190 950±70 590±190a (AU) 67.8 39.6 45.7 43.2 488 43.1 39.4 39.3e 0.44 0.25 0.15 0.19 0.84 0.04 0.22 0.25i (deg) 44.0 17.1 29.0 28.2 11.9 8.00 20.5 19.7r (AU) 96.8 31.2 52.0 51.1 88.5 43.3 47.8 42.1H -1.2 -1.0 -0.3 0.3 1.6 2.7 2.3 3.4surface composition CH4+? CH4+CO+N2 CH4+C2H6 H2O CH4+N2 H2O + ? H2O ?

albedo (%) 86±7 50-65 80+10

−20∼73 15-30 9±3 20±3 15+15

−6

mass (1020 kg) 166±2 130.5±0.6 - 42±1 - - 9±1 -

density (g cm−3) 2.3±0.3 2.03±0.06 - ∼2.6 - - 1.9±0.4 -satellite frac. brightness (%) 0.4 18,.018,.015 - 5.9,1.5 - 0.6 8 -satellite period (days) 15.8 6.4,38.2,24.8 - 49.1,? ? 9.8 -additional sat. limit (%) 0.04 0.001 0.01 0.5 0.2 0.2 0.1 0.5

Note.—References for all data can be found throughout the text.

16